From: Kevin M. Woods (kev@xxxxxxx)
Date: Mon May 22 2000 - 07:19:45 GMT-3
//
// AREA1/VLAN1
// --------------------
// |e0 |e0
// (R1) (R2)
// |s0 |s0
// | AREA0/FR |
// | |
// ------------------
// |s0.1 point to multipoint
// (R3)
//
// In this topology, If I have to send the traffic over
// R1 to VLAN1 always and sent over R2 only in case R1
// fails.. How do I do it. I have tried to give high IP
// OSPF COST at the s0 of R1,s0.1 on R1 and e0 of R1, but
// did not help. Always the traffic is going over only
// R2. I have done interarea summarisation over R1 and R2
// between area 1 and area 0. I have 'no ip route-cache'
// over all the interfaces. The problem is when I shut
// the interface e0 of R2 to check if the traffic now
// takes the route through R1 to reach VLAN1, it does not
// do that. The packets still try to go over to R2 and
// get lost.
It sounds like you wanted traffic destined to VLAN1 to traverse the VC
to R1 only. In that case you should increase the cost on the e0 of R2
and not R1. (Increasing the cost of the serial interfaces on R1 or R2
has no effect for what you were trying to do.)
How many networks did you summarize in VLAN1? If R2 still has a route
to any portion of the summary then it will still advertise it even
though you shutdown its e0. And since you increased the OSPF cost for
e0 on R1, R3 will still prefer R2 for getting to VLAN1.
// I am confused over the operation of 2 ABRs connecting
// same NON-Backbone area (area1) to Backbone area
// (area0).
//
// Are there any special considerations for interarea
// summarisation when you are using 2 ABRs to connect to
// area O.
Yes...you should add a summary route to Null0 to avoid routing loops in
certain situations.
// Any good place to read about these kinds of issues.
//
// Thakns in advance.
//
// Adwait
Kevin
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 08:23:30 GMT-3